Flow Deflectors for Preventing Stilling Basin

Abrasion Damage
by Leslie Hanna, U.S. Department of Interior, Bureau of Reclamation, Denver,Colorado

Abstract

Research sponsored by the U.S. Bureau of Reclamation’s (Reclamation) Science and

Technology Program and conducted by the Water Resources Research Laboratory in
Denver has demonstrated that flow deflectors can be used to mitigate abrasion damage
commonly experienced by Reclamation Type II stilling basins and other standard basins
of similar design. This will increase the life of the basins and reduce or eliminate the need
for costly repairs.

Introduction

Stilling basin abrasion damage is a widespread problem for river outlet works at dam sites
throughout the United States. Abrasion damage occurs when materials, such as sand,
gravel, or rock, are carried into the basin by a recirculating flow pattern produced over the
basin end sill during normal operation of a hydraulic jump energy dissipation basin
(figure 1). Once materials are in the basin, turbulent flow continually moves the materials
against the concrete surface, causing severe damage. Often abrasion damage has
occurred to the extent that reinforcing bars are exposed; then when repairs are made many
basins experience the same damage again within one or two operating seasons. Research
conducted by Reclamation’s Water Resources Research Laboratory (WRRL) in Denver has
demonstrated that the installation of flow deflectors can improve flow distribution, thus
minimizing or eliminating the potential for materials to be carried into stilling basins
(figure 2). This can increase the life of the basins and reduce necessary repairs.

Figure 1. Recirculating flow pattern occurs during normal

operation of hydraulic jump energy dissipation basin.
1
 Figure 2. Desired flow pattern with flow deflector installed.
The Mason Dam outlet works stilling basin, a typical Reclamation Type II basin with a long
history of abrasion damage, was determined to be an excellent candidate for a field
demonstration of this technology. Mason Dam is located on the Powder River in Baker
County, Oregon, approximately 17 miles southwest of the city of Baker. The dam was
constructed for irrigation, and for maintaining minimum flow in the Powder River. The dam
is a 173 ft high zoned earthfill embankment with a crest length of 895 ft. Reclamation owns
Mason Dam; however, the Baker Valley Irrigation District (BVID) operates and maintains
the facility under contract with Reclamation.

A physical model, constructed in the WRRL, was used to design a flow deflector for the
basin, and a field evaluation was conducted at the prototype facility to verify the
effectiveness of the design and to develop information for widespread application (a patent
is pending on this technology).

The Model

A 1:7 geometric scale was used to model the Mason Dam outlet works stilling basin.
Froude scale similitude was used to establish the kinematic relationship between model
and prototype because hydraulic performance depends predominantly on gravitational and
inertial forces. Froude scale similitude produces the following relationships between the
model and the prototype:

Length ratio Lr = 1:7

Velocity ratio Vr = Lr1/2 = 1:2.65

Discharge ratio Qr = Lr5/2 = 1:130

The physical model was used to investigate hydraulic conditions in the Mason Dam stilling
basin and to study the effect of deflector angle and position on flow patterns over the basin
end sill (figures 3 and 4).

2
Prototype features modeled included:

1) The two 33-in by 33-in high pressure

regulating gates and upstream bifurcation.

2) The 17 ft wide hydraulic jump twin bay stilling

basin with 2:1 sloping chutes, and dentated end
sill.

3) Approximately 75 ft of topography downstream

from the basin, constructed on a 5:1 slope with
moveable bed material.

Velocities were measured with a SonTek acoustic

doppler velocimeter (ADV) probe and were measured at
the downstream end of the basin at its centerline.
Tailwater elevation was set for each flow condition
tested, using tailwater data obtained during Mason Dam
outlet works operations. The deflector was modeled Figure 3. Looking upstream at
with a flat section of sheet metal spanning the 17 ft wide stilling basin model with ADV probe
basin and mounted on guides attached to the basin and deflector installed near end of
sidewalls, to allow vertical movement of the deflector basin.
within the basin.

Model Study Investigations

Model investigations were conducted to evaluate hydraulic conditions in the stilling basin
and downstream apron area for the range of operating conditions expected in the prototype.
The actual flow conditions
tested are listed in Table 1.
Velocity data and dye streak
data were collected and
analyzed to define basin
performance. This data was
used to determine the most
effective deflector angle and
the best lateral and vertical
locations within the basin.
Although investigations were
conducted up to the maximum
possible discharge of 870 ft3/s
(100% gate opening at
maximum reservoir), the Figure 4 Looking through the plexiglass sidewall of the
optimum deflector design was model operating at 40% gate opening.

3
based only on discharges up to 575 ft3/s (60% gate opening at maximum reservoir). This
is because Mason Dam’s standing operating procedures (SOP) limit outlet works discharges
to the maximum downstream river channel capacity of 500 ft3/s.

Velocities were measured at numerous locations within and downstream from the stilling
basin to map out resulting hydraulic flow patterns for each discharge tested. Early
investigations showed that average velocities measured at the end of the basin, at its
centerline, and 5.25 inches above the invert elevation provide a good representation of the
bottom velocities that carry materials into the basin (all dimensions and measurements
reported here are scaled to prototype dimensions). As a result, this location was used for
determining deflector performance. Average bottom velocities measured at this location
without a deflector ranged from -0.4 ft/s to -0.8 ft/s, with maximum upstream velocities in the
range of -2.0 ft/s to -3.0 ft/s (Negative velocities indicate flow is upstream into the basin).

Table 1. Prototype flow conditions tested in model.

Gate Opening Prototype Discharge Tailwater Depth (ft)
(%) Corresponding to
Maximum Reservoir
Elevation (ft3/s)
20 230 18.2
40 420 18.8
60 575 19.5
80 735 20.0
100 870 20.7

Model Study Results

Optimal Positioning and Size

Tests were initially conducted at 40 and 60 percent gate openings only, since these
conditions produced the strongest upstream bottom velocities adjacent to the riprap apron,
within the maximum operating range specified by the Mason Dam SOP. Four different
parameters were investigated to determine what criteria would produce best performance (all
parameters are referenced from the bottom downstream edge of the deflector):

! Vertical Elevation - Deflector elevation was varied from 4 ft to 15 ft above the

elevation of the basin floor (elevation 3889 ft).

! Lateral location - This is the distance from the downstream end of the stilling
basin (defined as the downstream end of the basin sidewalls) to the deflector.
Lateral locations were varied from 0 ft to 14 ft.

4
 ! Angle - Deflector angle was varied from 40 degrees to 90 degrees referenced
from the horizontal plane as shown in figure 2.

! Size - The height of the deflector was tested at 3 ft, 4 ft, and 5 ft.

The best performance, as determined by average bottom velocities measured at the

downstream end of the basin, occurred with a 5 ft deflector mounted 5 ft upstream from the
end of the basin at elevation 3900 ft (11 ft above basin floor), and angled at 90 degrees.

Overall Performance
After the optimal design parameters were set, it was important to look at deflector
performance with the basin operating throughout the full range of possible discharges up to
the maximum flow at 100% gate opening, in case unusual circumstances should require
releases above those normally allowed while the deflector is place. Table 2 shows the
average bottom velocities measured without a deflector compared with those measured with
the deflector set into optimal position.

Table 2. Basin performance with and without deflector.

Average prototype velocity measured in model at end of
basin with and without Deflector (ft/s)
Gate No Deflector Optimal Deflector at 3900 ft
Opening and angled at 90 degrees
(%)
20 -0.44 1.3
40 -0.73 1.8
60 -0.82 1.4
80 -0.88 -0.5
100 -0.69 -0.2

The table shows that with the deflector in place, performance at gate openings ranging from
20% to 60% is very good. Average velocities for this range of discharges are greater than
1.0 ft/s and are in the downstream direction. For gate openings of 80% and 100%,
performance is reduced significantly, however, performance is still improved over having no
deflector. Therefore, the deflector design was determined acceptable over the full range of
possible discharges.

In addition, if it were desired to obtain optimal performance throughout the full range of
possible operations, figure 5 shows that performance at higher discharges can be

5
significantly improved by moving the deflector to a lower elevation. This could be
accomplished with a mobile deflector supported on guides to allow a range of vertical
positions for operations at high and low discharges.

Average Velocity Measured at Downstream End of Basin (ft/s)

2.5

20% Gate
40% Gate
2
60% Gate
80% Gate
1.5 100% Gate

0.5

-0.5

-1

-1.5
No Deflector 3895.08 3897.42 3899.75 3902.08 3904.42
3893.92 3896.25 3898.58 3900.92 3903.25

Deflector Elevation (ft)

Figure 5. Average prototype bottom velocity measured in the

model vs. deflector elevation for each gate opening tested with
deflector angled at 80 degrees.

Hydraulically Self-Cleaning Operations

Model investigations showed that without a deflector, materials can be flushed from the basin
throughout the range of operations tested, due to the nature of the flow occurring within the
basin. This phenomenon occurs because turbulence within the basin periodically tosses
materials high enough into the water column to be caught and subsequently carried out by
the main jet exiting the basin. However, these suspended materials often hit their fall velocity
as they are exiting the basin and are deposited back onto the basin end sill; thereby making
them vulnerable to being carried right back into the basin by the upstream current. As a
result, for a large range of discharges, although materials are flushed out, the inflow of
materials is constant, thereby producing significant abrasion damage.

With the optimal deflector design in place, the model demonstrated that materials continue
to be flushed from the basin throughout the range of discharges tested. However, the
upstream component of velocity at the end of the basin is no longer strong enough to carry
a significant amount of material back into the basin. Therefore, the source for abrasion
damage can be eliminated during operations up to the maximum allowable discharge, and

6
the basin becomes hydraulically self-cleaning. The range of sizes of materials that can be
flushed from the basin will depend on outlet works operations and will be determined more
precisely in future studies.

Field Evaluation

The final prototype deflector for Mason Dam was designed with a set of guides that would
allow the deflector to be manually adjusted in angle and elevation for testing purposes. The
contract for the fabrication of the flow deflector was awarded to Prime Machine Inc. for
$27,000 including delivery to the site. In October of 2002 the flow deflector was delivered
to Mason Dam and installed by the Baker Valley Irrigation District (figure 6). In addition,
basin abrasion damage was repaired with new concrete at the time the flow deflector was
installed. The deflector was set to optimal position as determined from the model study
before seasonal operations began in April of 2003.

In August 2003, after nearly 5 months

of basin operations with the deflector
in place, a field evaluation and dive
inspection were conducted to verify
the effectiveness of the deflector.
Divers conducting the initial
underwater inspection noted that the
new concrete was very smooth and in
excellent condition, with no signs of
any erosion or wear. Although it may
be a little early to make any
conclusive statements, abrasion
damage is usually evident within one
to two operating seasons after repairs
have been made.

In addition, divers installed an

Acoustic Doppler Profiler (ADP) probe
at the downstream end of the basin to
measure exit velocities. The deflector
was raised above the water surface
and basin exit velocities were
measured at each 10% increment for
outlet works operations ranging from
10% gate opening up to 60% gate
opening. Then the same
measurements were repeated with
the deflector lowered to optimal
position, with bottom elevation at Figure 6. Deflector installation at Mason Dam outlet
3900 ft and angled at 90 degrees. works stilling basin in October 2002.

7
Table 3 shows the discharge tested at Mason Dam compared with the discharge tested in
the model for the same gate opening. The difference in values is because the model study
discharge was based on maximum reservoir elevation, and the reservoir was about 70 ft
below that level at the time tests were conducted at Mason Dam.

Table 3. Prototype discharges tested in the Model and at Mason Dam.

Gate Opening Prototype Discharge tested in Prototype Discharge tested at
(%) Model - Corresponding to Mason Dam at Low Reservoir
Maximum Reservoir (Elevation 4005 ft)
(Elevation 4075 ft) (ft3/s)
(ft3/s)
10 N/A 85
20 230 163
30 N/A 250
40 420 330
50 N/A 400
60 575 500

Figure 7 shows the average prototype velocities exiting the basin, measured at elevation
3891 ft (2 ft above the basin floor elevation) for each gate opening tested, with and without
a deflector. The figure shows significant improvement in the flow conditions at the
downstream end of the basin with the deflector lowered into optimal position for gate
operations from 10% to 30% gate opening. Average prototype velocities are greater than
0.75 ft/s and have changed from upstream in direction to downstream with the deflector in
place. However, for gate operations ranging from 40% to 60% gate opening, prototype
velocities measured were inconclusive due to limitations of the ADP probe to accurately
measure velocities when large quantities of air are entrained in the flow. The deflector was
designed to redirect the main jet exiting the basin down toward the basin end sill. Therefore
at high discharges, when the jet is highly aerated, entrained air was also redirected
downward towards the endsill where the ADP probe was located. As a result, accurate
velocity measurements were not possible at the higher discharges.

Figure 8 compares model and prototype average exit velocities, measured at elevation
3891 ft for each gate opening tested, with and without a deflector. The ADV probe used in
the model study was not as sensitive to air concentration, therefore, velocity measurements
were possible for all configurations tested. Although model and prototype discharges are not
identical (due to low reservoir elevation during prototype testing) figure 8 shows a strong
correlation between model and prototype velocities measured at the same location for the
same gate openings. Therefore, it would be reasonable to assume, with the field verified
data already acquired, that the velocities measured in the model for gate openings ranging
from 40% to 60% (with the deflector in place) are also a reasonable representation of
prototype flow conditions.

8
 2

1.5

Average Velocity (ft/s)

0.5

-0.5

-1

-1.5

P rototype - N o D eflector
-2
P rototype - D eflector at 3900 ft

-2.5

-3
0 100 200 300 400 500 600
3
O utlet W orks Discharge (ft /s)

Figure 7. Average prototype velocity measured at downstream

end of stilling basin at Mason Dam at an elevation 2 ft above the
basin floor, with and without deflector.

1 .5

1
Average Velocity (ft/s)

0 .5

-0 .5

-1

-1 .5 M ason Prototype No D eflector

-2 M ason Prototype W ith Deflector @ 3900 ft

-2 .5 M odel No deflector

M odel W ith D eflector @ 3900 ft

-3
0 100 200 300 400 500 600
D ischarge (ft 3 /s)

Figure 8 Average prototype velocities measured in the model

and the prototype at downstream end of stilling basin at an
elevation 2 ft above the basin floor, with and without deflector.
Revised model data included.

9
A second field evaluation and dive inspection are scheduled for the Mason Dam outlet works
stilling basin for August 2004. At that time, an ADV probe will be used (in addition to the ADP
probe) to measure velocities at the end of the basin to further verify model data and to
conduct additional testing.

Generalizing Deflector Design for Widespread Applications

The model investigations and field evaluation were used to develop a method for generalizing
deflector design for Reclamation Type II and similar basins, based on velocity profiles
measured at the end of the basin before a deflector is installed. For future installations,
velocity data measured on-site can be used to determine the optimal design and location for
a deflector for a specific basin.

At facilities where the outlet works is operated up to its full capacity of 100 percent gate
opening, two different options can be considered:

! One option is to design the deflector to be effective for the most predominant
range of basin operations. This would mean that when the basin was operated
outside the deflector design range, some materials may be drawn into the
basin. In this case, it would be recommended that the basin be operated within
the designated design range periodically, to purge materials from the basin.

! A second option would be to design a moveable deflector supported on guides

so that deflector elevation could be changed for different ranges of operations.
In most cases this would require only two positions.

Implementation of either option would significantly reduce the amount of damage caused by
abrasion and the costs associated with basin repairs.

References

Dodge, Russ, “Hydraulic Model Study of Taylor Draw Dam Outlet Works”, Water Resources
Research Laboratory, Bureau of Reclamation, R-96-09, December 1996.

10